I don't have a definitive answer. I think teachers and lecturers are really well trained in spotting AI content. I think my lecturers knew when someone used AI. And a woman I dated was teacher and she was pretty good in spotting AI texts. Though, when we started texting I used AI feedback and honestly I think she never noticed anything. I also adapted it to my style.
Personally, I have the feeling the hyphen/em dash used by chatbots (not sure whether other chatbots use it). My theory is chatGPT uses this as obvious marker for AI texts. And that people can feel smarter when they delete the hyphen/em dash to make it invisible that they used AI. But of course that's bullshit. You can also ask an AI chatbot whether a text they created sounds AI generated. AI detectors often can detect it while the AI chatbots denies that fact. I use dating apps. And we and my friends joke when we see profiles where it is either obvious they used AI or where we speculate about it. I think some people have become so lazy that they don't even delete the hyphen/em dash. Maybe they even don't want to hide it. Personally, I told my friends a theory about an app where one could date in the future. The AI agents of different people make connections. And this is how the dating future will look like. That could be dystopian. I think it would be an invasion of privacy and one reason why no one would like to use it. Because there are not enough filters which information one is willing to share or put into the profile. To come back to my point. Maybe some people consider it completely normal to use AI chatbots for dating apps. It is more personalized then and they use it to express their character. I had to chuckle a woman used a text in quotation marks for an answer in her profile. A classical beginner's mistake. The statement was obvious AI slope. Sometimes it is easier to spot AI content. Maybe that's an illusion. But I think the texts chatGPT generates in my nativer language are harder to detect. It is adapted to my German voice quite impressively. It is still too smooth. It takes the edges of my humor. So it is less controversial and my texts scream "I hate myself" a little bit less.
Honestly, I think on English using AI would be way too easy to detect on here. At least for the people who know my writing style. My friends criticize me for my poor English language skills on here. My sentence structure is far too German. And chatGPT also concluded my English must have deteriorated a lot since I left school. I had very very good English marks. But honestly on here I also don't put that much effort in it. Writing feels cathartic. Trying to strcture it and correcting my mistakes. It feels like a prison. I tend to extreme self-observation online. And when I start revising my texts I have a hard time to stop it. If I start this perfectionism on here writing will stop being a coping skill and it will become work. This is something I realized when I worked on a project for a more sophisticated text than usually.
I know this has nothing to do with the core idea of this thread. My rambling is heavily criticized when I ask for AI feedback. I gave college another shot last year for like one month. I almost broke down after a short time. There was a theory seminar. And I tried to impress the lecturer with my reading. This is pathological and I perceive this as very pathetic. Fake it till you make it. But this often works at the universities and in academia. But I digress. He really liked my first essay that I handed in. I was extremely paranoid about using AI. I sort of finished the text with the help of AI. And I completely rewrote the text and I had a completely new version because I threw all AI text away. I think I had the best mark from the people who handed in a text. But it was only a 1,7. (scale 1.0-5.0) Though, he complimented me that he was very impressed by my writing style (on German). Honestly, I think I had to learn a lot about academic writing to get better grades. But with my mental health I will never go back. The second text I handed it. It was about a text from very popular German philosopher. And I often have a hard time to follow him. I know ZIziek likes him. But I don't like the style of this intellectual and the topics he writes about. I often don't feel a connection. But I had to write about him. And this time I had no more energy left to rewrite the text. And I used AI content. I got the same grade. Again the best mark from the people who handed something in. And I used an AI detector on texts of other people. And one woman handed in a text that was fully AI generated like 100% and it sounded extremely generic. And it sounded like she took the first result and ran with it. I didn't do that. I put a lot of effort in it. But I think I also started to develop an addiction to AI chatbots when it comes to academic texts. I realized retrospectively in many ways how I used them them this was counterproductive. And I would not use them in this way again. ChatGPT compared my German to the German of Adorno by the way. Damn I hate sycopaphancy but I wish this was true. There is a funny anecdote. I barely write on German on here. I think my German has strong idiosyncracies. And I think it would make it too easy to spot me on other platforms. There was something I chuckled a lot about. In one German thread I wrote like one sentence in German. And then someone came and said. "Don't trust this member. His German sounds very sus". Something like that. I had like extremely good marks in German. And someone else replied "With this post count and the fact he is online for quite some years it is far more likely that you are a fed." I paraphrase it I don't know the literal quotes anymore.
I feel pretty stupid that I have a hard time to decipher whether a text was written by a chatbot or not. And I think the fact that chatGPT adapts my writing skill makes it harder for me to see the difference. But on SaSu on English a text without my usual mistakes that would it obvious. I sometimes used AI for threads on here but I usually I am transparent about it.
Recently in a newspaper I read a pretty good aritlce about the writing of AI chatbots.
I try to summarize it: AI has a very strong use of pathos. Certain idioms and phrases, repitions of content, Sometimes the texts repeate themselves over and over. (I wonder about that because chatGPT says that about my human written texts too). Lack of details, weak usage of sources. Subtle cracks in structure and chronology (I can see where they are coming from). I can also agree with the following. Different parts of the content clash without making necessary connections between them. Generalizations and platitudes about topics where there is a lack in the training data. AI chatbots are not able to read between the lines. This is something I thought a lot about. Because I use them for feedback on dating chats. Whether they simply can't see the subtle nuances that I see. Most of the time the subtle nuances I notice end up to be pathological paranoia though. Lol. The texts lack a human signature/handwriting. In a different text I read people describe AI texts as soulless. WIth sometimes really inappropriate and horrible comparisons when you think more about them. I read this because someone might use an AI chatbot to write about the holocaust.
Eventually I quit college. It was horrible because I was always so scared to be exposed as fraud. And I always was so anxious to have disappointed my Marxist lecturer. Lol. That sorrow is pathological. I had such delusions in the past. But I think he was actually really pissed that someone who postured himself as Zizek fan and argumented against capitalism and tech billionaires used AI chatbot for their writing. I can remember it was extremely uncomfortable for me that he noticed it. In a subtle way he confronted me with it. And he admitted he was paranoid himself about students using AI and that he hopes so much that this bubble bursts eventually. My friends in real life hope the same. I quit the seminar and college eventually. But even if I didn't use AI I would have quit. I am in a horrible shape and college made it so much worse. I think though the way I used AI didn't make my writing better. I should have invested the time in research of primary sources instead. I am way too susceptible for these feedback loops. I also have the personal traits and the living circumstances that make me prone to be an AI heavy user. Though he gave me again a 1,7 as grade. But I felt guilty and horrible. These traits and their consequences would be worth to be elaborated in another essay. Lol. ChatGPT will probably let me know this tomorrow. I have implemented a task to give me everyday one idea for a potential essay that I could write. But bro that's way too much work. And honestly I think it cannot really compete with my creativity when it comes to thread ideas for this forum.
In your opinion? What makes a text suspicious? How do you spot AI content?
Personally, I have the feeling the hyphen/em dash used by chatbots (not sure whether other chatbots use it). My theory is chatGPT uses this as obvious marker for AI texts. And that people can feel smarter when they delete the hyphen/em dash to make it invisible that they used AI. But of course that's bullshit. You can also ask an AI chatbot whether a text they created sounds AI generated. AI detectors often can detect it while the AI chatbots denies that fact. I use dating apps. And we and my friends joke when we see profiles where it is either obvious they used AI or where we speculate about it. I think some people have become so lazy that they don't even delete the hyphen/em dash. Maybe they even don't want to hide it. Personally, I told my friends a theory about an app where one could date in the future. The AI agents of different people make connections. And this is how the dating future will look like. That could be dystopian. I think it would be an invasion of privacy and one reason why no one would like to use it. Because there are not enough filters which information one is willing to share or put into the profile. To come back to my point. Maybe some people consider it completely normal to use AI chatbots for dating apps. It is more personalized then and they use it to express their character. I had to chuckle a woman used a text in quotation marks for an answer in her profile. A classical beginner's mistake. The statement was obvious AI slope. Sometimes it is easier to spot AI content. Maybe that's an illusion. But I think the texts chatGPT generates in my nativer language are harder to detect. It is adapted to my German voice quite impressively. It is still too smooth. It takes the edges of my humor. So it is less controversial and my texts scream "I hate myself" a little bit less.
Honestly, I think on English using AI would be way too easy to detect on here. At least for the people who know my writing style. My friends criticize me for my poor English language skills on here. My sentence structure is far too German. And chatGPT also concluded my English must have deteriorated a lot since I left school. I had very very good English marks. But honestly on here I also don't put that much effort in it. Writing feels cathartic. Trying to strcture it and correcting my mistakes. It feels like a prison. I tend to extreme self-observation online. And when I start revising my texts I have a hard time to stop it. If I start this perfectionism on here writing will stop being a coping skill and it will become work. This is something I realized when I worked on a project for a more sophisticated text than usually.
I know this has nothing to do with the core idea of this thread. My rambling is heavily criticized when I ask for AI feedback. I gave college another shot last year for like one month. I almost broke down after a short time. There was a theory seminar. And I tried to impress the lecturer with my reading. This is pathological and I perceive this as very pathetic. Fake it till you make it. But this often works at the universities and in academia. But I digress. He really liked my first essay that I handed in. I was extremely paranoid about using AI. I sort of finished the text with the help of AI. And I completely rewrote the text and I had a completely new version because I threw all AI text away. I think I had the best mark from the people who handed in a text. But it was only a 1,7. (scale 1.0-5.0) Though, he complimented me that he was very impressed by my writing style (on German). Honestly, I think I had to learn a lot about academic writing to get better grades. But with my mental health I will never go back. The second text I handed it. It was about a text from very popular German philosopher. And I often have a hard time to follow him. I know ZIziek likes him. But I don't like the style of this intellectual and the topics he writes about. I often don't feel a connection. But I had to write about him. And this time I had no more energy left to rewrite the text. And I used AI content. I got the same grade. Again the best mark from the people who handed something in. And I used an AI detector on texts of other people. And one woman handed in a text that was fully AI generated like 100% and it sounded extremely generic. And it sounded like she took the first result and ran with it. I didn't do that. I put a lot of effort in it. But I think I also started to develop an addiction to AI chatbots when it comes to academic texts. I realized retrospectively in many ways how I used them them this was counterproductive. And I would not use them in this way again. ChatGPT compared my German to the German of Adorno by the way. Damn I hate sycopaphancy but I wish this was true. There is a funny anecdote. I barely write on German on here. I think my German has strong idiosyncracies. And I think it would make it too easy to spot me on other platforms. There was something I chuckled a lot about. In one German thread I wrote like one sentence in German. And then someone came and said. "Don't trust this member. His German sounds very sus". Something like that. I had like extremely good marks in German. And someone else replied "With this post count and the fact he is online for quite some years it is far more likely that you are a fed." I paraphrase it I don't know the literal quotes anymore.
I feel pretty stupid that I have a hard time to decipher whether a text was written by a chatbot or not. And I think the fact that chatGPT adapts my writing skill makes it harder for me to see the difference. But on SaSu on English a text without my usual mistakes that would it obvious. I sometimes used AI for threads on here but I usually I am transparent about it.
Recently in a newspaper I read a pretty good aritlce about the writing of AI chatbots.
I try to summarize it: AI has a very strong use of pathos. Certain idioms and phrases, repitions of content, Sometimes the texts repeate themselves over and over. (I wonder about that because chatGPT says that about my human written texts too). Lack of details, weak usage of sources. Subtle cracks in structure and chronology (I can see where they are coming from). I can also agree with the following. Different parts of the content clash without making necessary connections between them. Generalizations and platitudes about topics where there is a lack in the training data. AI chatbots are not able to read between the lines. This is something I thought a lot about. Because I use them for feedback on dating chats. Whether they simply can't see the subtle nuances that I see. Most of the time the subtle nuances I notice end up to be pathological paranoia though. Lol. The texts lack a human signature/handwriting. In a different text I read people describe AI texts as soulless. WIth sometimes really inappropriate and horrible comparisons when you think more about them. I read this because someone might use an AI chatbot to write about the holocaust.
Eventually I quit college. It was horrible because I was always so scared to be exposed as fraud. And I always was so anxious to have disappointed my Marxist lecturer. Lol. That sorrow is pathological. I had such delusions in the past. But I think he was actually really pissed that someone who postured himself as Zizek fan and argumented against capitalism and tech billionaires used AI chatbot for their writing. I can remember it was extremely uncomfortable for me that he noticed it. In a subtle way he confronted me with it. And he admitted he was paranoid himself about students using AI and that he hopes so much that this bubble bursts eventually. My friends in real life hope the same. I quit the seminar and college eventually. But even if I didn't use AI I would have quit. I am in a horrible shape and college made it so much worse. I think though the way I used AI didn't make my writing better. I should have invested the time in research of primary sources instead. I am way too susceptible for these feedback loops. I also have the personal traits and the living circumstances that make me prone to be an AI heavy user. Though he gave me again a 1,7 as grade. But I felt guilty and horrible. These traits and their consequences would be worth to be elaborated in another essay. Lol. ChatGPT will probably let me know this tomorrow. I have implemented a task to give me everyday one idea for a potential essay that I could write. But bro that's way too much work. And honestly I think it cannot really compete with my creativity when it comes to thread ideas for this forum.
In your opinion? What makes a text suspicious? How do you spot AI content?
Last edited:
